- Email: [email protected]
Expression of Amiloride-Sensitive Sodium Channel: A Strategy for the Coexpression of Multimeric Membrane Protein in Sf9 Insect Cells
Analytical Biochemistry 286, 206 –213 (2000) doi:10.1006/abio.2000.4861, available online at http://www.idealibrary.com on
Expression of Amiloride-Sensitive Sodium Channel: A Strategy for the Coexpression of Multimeric Membrane Protein in Sf9 Insect Cells U. Subrahmanyeswara Rao, 1 A. Mehdi, and R. E. Steimle Department of Biochemistry and Molecular Biology, University of Nebraska Medical Center, Omaha, Nebraska 68198
Received February 29, 2000
The amiloride-sensitive epithelial sodium channel (ENaC) mediates Na ⴙ reabsorption in many epithelial tissues including the distal nephron, colon, lung, and secretory glands and plays an important role in pathophysiology of hypertension and cystic fibrosis. The ENaC is a multimeric integral membrane protein formed by the association of highly homologous,␣-, -, and ␥-ENaC subunits. Here we explored the Sf9 insect cell– baculovirus expression system as a source to obtain high yields of recombinant ENaC for functional and structural studies. Although this expression system is widely used, coexpression of ENaC subunits could not be accomplished by the conventional procedures. We thus developed a protocol in which the ␣and ␥-ENaC cDNA’s were first fused individually with polyhedrin promoters at their 5ⴕ-ends and then inserted in the multiple cloning sites of pVL1393 transfer vector carrying the -ENaC cDNA. Utilizing this transfer vector, a recombinant baculovirus carrying all of the three ENaC cDNA’s was prepared. Infection of Sf9 insect cells with this recombinant baculovirus resulted in the expression all of the three ENaC subunits in high yield. Planar lipid bilayer reconstitution procedure revealed the presence of ⬃6 pS sodium channels that are amiloride-sensitive. The results presented point out certain underlying rules for the expression of multiple genes in Sf9 cells, which may be useful in the expression other multimeric proteins and in the studies of protein–protein interactions as well. © 2000 Academic Press
1 To whom correspondence should be addressed at Department of Biochemistry and Molecular Biology, Campus 98-4525, University of Nebraska Medical Center, Omaha, NE 68198-4525. Fax: 402-5596650. E-mail: [email protected].
206
Development of heterologous protein expression systems such as Escherichia coli, yeast, mammalian, and Sf9 cells has revolutionized our understanding of detailed structure and function of numerous proteins (1– 7). In many instances, the expressed recombinant proteins are monomeric and function independently without the association with other proteins (8, 9). However, the scenario is different for multimeric proteins, which require coexpression of all the required subunits to elicit their functional characteristics (10, 11). Similarly, although certain proteins may independently be expressed, the biological effects of the expressed protein can only be ascertained by their interactions with other proteins. Production of biologically active multimeric recombinant proteins has been difficult and often hampered by the lack of simple procedures for simultaneous expression of all the required subunits in each of the expressing host cell. While expression of such multimeric proteins in yeast can be accomplished easily due to the availability of multiple auxotrophic yeast strains and vectors (3), the formidable and inefficient yeast cell-wall breakage procedures greatly reduce the recovery of the expressed proteins to unacceptable levels for further structural or functional analyses. Recently, the Sf9 cell– baculovirus (BV) 2 expression system has become an important tool for the largescale production of foreign proteins (12–16). Despite the fact that functional multimeric proteins were also expressed in Sf9 cells, expression of subunits of these protein complexes were interestingly accomplished by coinfection with multiple BVs each harboring the required subunit cDNA. The issues, for example, whether exposure with multiple recombinant BVs results in the infection of each Sf9 cell with all of the viruses used, and whether the expressed subunits to 2 Abbreviations used: BV, baculovirus; ENaC, epithelial sodium channel; PBS, phosphate-buffered saline.
0003-2697/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
SUBUNIT COEXPRESSION OF MULTIMERIC PROTEIN IN Sf9 CELLS
what extent are in the multimeric state, are however unclear from the literature survey. It is thus desirable to prepare a single recombinant BV harboring all of the required cDNA’s for the production of fully functional multimeric proteins in high yields. Although strategies for the transfer of multiple cDNA’s into the baculoviral genome have been formulated (12–17), such procedures for the production of biologically important functional multimeric proteins in Sf9 cells were not reported in the literature. The long-term goal of our laboratory is to understand the amiloride-sensitive epithelial sodium channel (ENaC)-mediated sodium transport. It has been established that fully functional ENaC is an integral membrane protein composed of three highly homologous subunits, termed ␣-, -, and ␥-ENaC (18, 19). To develop the Sf9 insect cell–BV expression system as a source of recombinant ENaC, we have examined several strategies to coexpress all of the three ENaC subunits in Sf9 cells. The results reported in this paper point out that coexpression of ENaC subunits in Sf9 cells is not a trivial matter, and the conventional strategy of infection of cells with multiple recombinant BVs was inefficient. Here we report a procedure that utilizes the most commonly used pVL1393 transfer vector to coexpress all of the three cDNA’s of ENaC in Sf9 cells in high amounts. The results of this study reveal that when two transcription units in the transfer vector are oppositely oriented, the expression of corresponding proteins was very low, probably due to positive supercoiling of DNA due to transcription from both ends. However, the expression is maximum when these transcription units are present tandemly in the baculoviral genome. These observations may be important in the considerations of coexpression of other multimeric proteins and in the studies of protein-protein interactions as well. MATERIALS AND METHODS
cDNA’s. The rat colon ENaC ␣-, -, and ␥-subunit cDNA’s (18, 19) were first subcloned under the control of polyhedrin promoter in the baculoviral transfer vector, pVL1393, and generated three transfer vectors, pVL1393/␣, pVL1393/ and pVL1393/␥, respectively. The ENaC cDNA’s from these vectors were used for subsequent subcloning into other baculoviral transfer vectors. Construction of baculoviral transfer vectors carrying ENaC cDNA’s and preparation of recombinant BVs. Standard molecular biological procedures of subcloning were used (20). Details of the construction of a variety of baculoviral transfer vectors carrying ENaC subunit cDNA’s are described under Results. The Sf9 cells were cotransfected with baculoviral transfer vectors and the BaculoGold DNA by the calcium phos-
207
phate coprecipitation method according to the protocol provided by the manufacturer (PharMingen, San Diego, CA) and prepared recombinant BVs. Baculoviral infections of Sf9 cells and preparation of membrane fraction. Sf9 cells were grown and infected with the recombinant BVs at a ratio of 1:5, and the total membrane fraction was prepared as described (21). SDS–PAGE and immunoblotting. Proteins in the membrane fractions were first separated on 7.5% acrylamide gels containing 0.1% sodium dodecyl sulfate and transferred onto the poly(vinylidene fluoride) membranes as described previously (21). Immunodetection of ENaC subunits was carried out by using ENaC subunit-specific polyclonal antibodies raised in rabbits in conjunction with a commercially available horseradish peroxidase-linked anti-rabbit secondary antibody (Amersham, IL). The peroxidase-labeled blots were developed by the enhanced chemiluminescence method, using the Amersham ECL Plus kit (21). Immunofluorescence microscopy. Recombinant BVinfected Sf9 cells were placed on microscope slides and fixed with neat glutaraldehyde followed by immersion in cold acetone and methanol. The slides were treated with BSA buffer (1% (w/v) bovine serum albumin dissolved in phosphate-buffered saline (PBS)) for 1 h at room temperature. The slides were then incubated with the ENaC-specific antibody diluted in PBS for 3 h, washed three times in BSA buffer, and then incubated with Alexa-coupled anti-rabbit antibody (Molecular Probes, Eugene, OR) prepared in PBS (1000:1 dilution) for 1 h at 4°C in dark. The slides were washed with PBS and fixed with antifade solution, and the Alexalabeled cells were detected under Zeiss LSM 410 confocal laser scanning microscope and the images were digitally recorded on a computer. Planar lipid bilayer reconstitution. Single channel analysis of the ENaC complex expressed by the 1393/ ␣(⫹)␥(⫹)-BV in Sf9 cell membranes was performed in planar lipid bilayers by the general procedures of Rosenberg and Chen (27) with modifications. Planar lipid bilayers were formed by painting 40 mg/ml solution of phospholipid (phosphatidylethanolamine:phosphatidylserine at a ratio of 3:1) in n-decane over a 200-m-diameter aperture separating cis and trans chambers. The cis compartment of the bilayer chamber contained 160 mM sodium aspartate, whereas, the trans chamber contained 60 mM sodium aspartate, in 10 mM Hepes, 0.5 mM CaCl 2, pH 7.4, buffer. Channel activity was monitored after the addition of Sf9 cell membranes to the cis chamber. Single channel currents collected at different voltage potentials were filtered at 100 Hz and analyzed by using pCLAMP 7.0 software (Axon Instruments).
208
RAO, MEHDI, AND STEIMLE
FIG. 1. Expression of ENaC subunits by multiple baculoviral infection. Sf9 cells were infected with a pool of ␣-,-, and ␥-BVs, and the membranes prepared were run on SDS–PAGE followed by electroblotting onto PVDF membranes (lanes 2). Membranes obtained from Sf9 cells infected with a single BV were run in parallel (lanes 1). Membranes from cells infected with BV carrying MDR1 cDNA were used as control (lanes 3). The immunoblots were developed with anti-␣, anti-, and anti-␥ antibodies. Arrows indicate the expected glycosylated and nonglycosylated subunits and their sizes.
Protein estimation. Protein in the membrane fractions was determined by the modified Lowry method, using bovine serum albumin as standard (22, 26). Materials. BaculoGold DNA and transfer vectors, pVL1393, pAcSG2, and pAcAB3 were obtained from PharMingen. The Sf9 culture media were obtained from GibcoBRL. RESULTS
Coexpression of ENaC subunits by infections with multiple recombinant BVs. To accomplish the coexpression of ␣-, -, and ␥-ENaC subunits in each Sf9 cell, the conventional procedure of infecting cells with multiple recombinant BVs was tested and the results are shown in Fig. 1. The Sf9 cells were infected with a pool of ␣-, -, and ␥-BVs and the amount of expression of each ENaC subunit in the membranes was compared with the corresponding subunit expressed in cells infected with a single BV. At least eight proteins with molecular masses in the range of ⬃97– 60 kDa were reactive with the anti-␣-ENaC antibody in membranes obtained from cells infected with ␣-BV (lane 1, anti-␣). The 97- and 80-kDa proteins are, respectively, the fully glycosylated and nonglycosylated forms of ␣-ENaC subunit, and the remaining protein bands represent proteolytic products (data not shown). Infection of cells with the pool of ␣-, -, and ␥-BVs also resulted in the expression of ⬃97- and 80-kDa ␣-ENaC subunits (lane 2, anti-␣), but the staining intensities were at least 20-fold lower than the amounts of ␣-ENaC expressed by the ␣- BV infection. Figure 1 (anti-) shows that cells infected with the -BV expressed ⬃97-kDa (glycosylated) and 70-kDa (nonglycosylated) -ENaC subunit forms (lane 1), the amounts of which were greatly reduced in cells infected with the pool of ␣-, -, and ␥-BVs (lane 2). Similarly, infection of Sf9 cells with a single ␥-BV resulted in the expression of ⬃80-kDa (glycosylated) and 75-kDa (nonglycosylated) ␥-ENaC subunits (Fig. 1, lane 1, anti-␥). However the staining intensities of these subunit forms were again greatly
reduced by coinfection of cells with the pool of ␣-, -, and ␥-BVs (Fig. 1, lane 2). We have varied the multiplicity of infection from 1:1 to 1:25 with essentially the same results (data not shown). These results suggested that coinfection of Sf9 cells with multiple BVs results in the reduced expression of each of the three ENaC subunits. To determine the reason for the reduced expression of ENaC subunits in cells infected with multiple BVs, Sf9 cells were infected with the pool of ␣-, -, and ␥-BVs and the ENaC subunits in these cells were visualized by immunofluorescence microscopy using antibodies specific to the three ENaC subunits. The results show (Fig. 2) that about 10% of the cells were fluorescent with each of the three ENaC subunit-specific antibodies, further corroborating the above immunoblotting analyses, in which the expression of these subunits was greatly reduced (Fig. 1). In contrast, nearly 100% of the cells were fluorescent when infected with any single BV (data not shown). These data suggest that cells once infected with one recombinant BV may not be further infected with any other BV. Alternatively, all of the cells were in fact infected with three ENaC BVs, but unable to produce detectable amounts of recombinant proteins. Coexpression of ENaC subunits via infection with a BV carrying three ENaC cDNA’s. To express the three ENaC subunits in each infected Sf9 cell, a BV
FIG. 2. Visualization of ENaC subunits in Sf9 cells infected with a pool of 3 ENaC BVs. Sf9 cells were infected with a pool of ␣-, -, and ␥-BVs for 3 days. The cells were fixed on glass slides and incubated with anti-␣, anti-, or anti-␥ antibody, followed by Alexa-labeled anti-rabbit secondary antibody, as described under Materials and Methods. (A) The cells as observed under bright field. The ENaC antibodies were visualized with Alexa-labeled secondary antibody by indirect immunofluorescence as observed under scanning microscope (B).
SUBUNIT COEXPRESSION OF MULTIMERIC PROTEIN IN Sf9 CELLS
FIG. 3. Construction of pAcSG3 vector as a vehicle for the expression of multiple ENaC subunits. (A–C) The construction of pAcSG3 transfer vector carrying three promoters and insertion of the three ENaC cDNA’s downstream to these promoters (see Results for details). (D) Membranes prepared from Sf9 cells infected with SG3/␣-, SG3/-, SG3/␥-, SG3/␣-, and SG3/␣␥-BVs were analyzed for the expression of ␣-, -, and ␥-ENaC subunits by immunoblotting using anti-␣, anti-, and anti-␥ ENaC antibodies. Membranes obtained from SG3/␣-BV-infected (anti-␣), SG3/-BV-infected (anti-), and SG3/␥-BV-infected (anti-␥) cells were loaded in lanes 1. Lanes 2 and 3 were loaded with membranes prepared from Sf9 cells infected with SG3/␣- and SG3/␣␥-BVs, respectively. Lanes 4 were loaded with control Sf9 cell membranes.
carrying all of the three ENaC subunit cDNA’s is necessary. Toward this end, a small baculoviral transfer vector, pAcSG3 vector containing three strong baculoviral promoters was prepared (Fig. 3) in the following way. Step 1: The commercially available ⬃10 kbp pAcAB3 baculoviral transfer vector containing three baculoviral promoters (one polyhedrin and two p10 promoters) was digested with NaeI and SnaBI and the released 1.26 kb DNA fragment that contained all of the three promoters was isolated (Fig. 3B). Step 2: A single polyhedrin promoter containing 5.5 kb baculoviral transfer vector, pAcSG2 (Fig. 3A), was digested with NaeI and SnaBI and the released 572-bp polyhedrin promoter- and the multiple cloning sites-containing fragment was discarded. Step 3: The 1.26 kb NaeI–SnaBI fragment isolated from Step 1 was ligated with the ⬃5-kb fragment isolated in Step 2 and generated a small 6.2-kb vector termed, pAcSG3, which contained one polyhedrin and two p10 promoters (Fig. 3C).
209
The ␣-ENaC cDNA was first subcloned at StuI site under the control of polyhedrin promoter, and prepared a baculoviral transfer vector, pAcSG3/␣. The -ENaC cDNA was then subcloned under the control of a p10 promoter at the SmaI site of the pAcSG3/␣ vector and generated a baculoviral vector, pAcSG3/␣. In a final step, the ␥-ENaC cDNA was subcloned under the control of the second p10 promoter at the CelII site and prepared a baculoviral vector, pAcSG3/␣␥. Using pAcSG3/␣, pAcSG3/␣, and pAcSG3/␣␥ vectors, three recombinant BVs termed SG3/␣-, SG3/␣-, and SG3/ ␣␥-BVs, which harbored, respectively, one, two, and three ENaC cDNA’s, were prepared. Total membrane fraction prepared from Sf9 cells infected with SG3/␣-, SG3/␣-, and SG3/␣␥-BVs were analyzed for the presence of ENaC subunits by immunoblotting procedure and the results are shown in Fig. 3D. Infection with SG3/␣-BV results in the expression of ␣ENaC subunit (Fig. 3, lane 1, anti-␣). Similarly, BVs prepared from the pAcSG3 vector carrying single -ENaC or ␥-ENaC cDNA, termed SG3/ (Fig. 3, lane 1, anti-) and SG3/␥ (Fig. 3, lane 1, anti-␥) allowed the expression of these subunits. However, cells infected with SG3/␣-BV (Fig. 3, lane 2) or SG3/␣␥-BV (lane 3) did not contain any detectable ␣-ENaC. There was no detectable -ENaC subunit expression in cells infected with SG3/␣- or SG3/␣␥-BVs (Fig. 3, lanes 2 and 3, anti-). Similarly, SG3/␣␥-BV infection did not result in any detectable ␥-ENaC subunit expression (Fig. 3, lane 2, anti-␥). These data suggest that although pAcSG3 vector allows high-yield expression of a single cDNA, it is not suitable for the expression of multiple cDNA’s in the Sf9 cells. High-yield expression of multiple proteins using single pVL1393 transfer vector. We have devised a new strategy to accomplish the goal of coexpression of ENaC subunits in Sf9 cells. In this procedure, the ENaC subunit cDNA’s were first ligated with polyhedrin promoters at their 5⬘-ends and generated individual transcription units which were then inserted in the multiple cloning sites of pVL1393 transfer vector. The resulting transfer vector thus contained multiple cDNA’s whose expressions were driven by polyhedrin promoters placed at their 5⬘-ends. Figure 4A shows schematically a pVL1393 transfer vector with two ENaC cDNA’s. As a first step, we have inserted the -ENaC cDNA in pVL1393 transfer vector under the control of polyhedrin promoter and prepared pVL1393/ -ENaC transfer vector. An ⬃3-kb DNA fragment comprising the entire polyhedrin promoter and the ␣-ENaC cDNA (PHDRN␣) was isolated from the pAcSG3/␣ after digestion with BamHI, and inserted at the BglII site of pVL1393/ENaC. Because BglII and BamHI cleaved ends are compatible, insertion of PHDRN␣ into pVL1393/-ENaC generated two trans-
210
RAO, MEHDI, AND STEIMLE
FIG. 4. Construction of pVL1393 transfer vector carrying two transcription units and protein expression. (A) The ␣-ENaC cDNA with polyhedrin promoter at the 5⬘-end (PHDRN␣) was prepared and then inserted downstream to the -ENaC cDNA in the pVL1393/ transfer vector. The two possible orientations of inserted PHDRN␣ are indicated. Membranes prepared from cells infected with 1393/ ␣(⫹)-BV (lanes 1), 1393/␣(⫺)-BV (lanes 2), and 1393/␥␣(⫺)-BV (lane 3) were run on SDS–PAGE followed by immunoblotting. The blots were developed with anti-␣ (B) and anti- C).
fer vectors, pVL1393//PHDRN␣(⫹) and pVL1393// PHDRN␣(⫺) in which the PHDRN␣ was inserted, respectively, in tandem and in reverse orientations, with respect to the orientation of -ENaC cDNA. The recombinant BVs prepared from these vectors are termed, respectively, 1393/␣(⫹)- and 1393/␣(⫺)-BVs. Expression of ␣- and -ENaC subunits in Sf9 cells infected with the 1393/␣(⫹)-BV was assessed by immunoblotting procedure and the results are shown in Figs. 4B and 4C. Cells infected with 1393/␣(⫹) BV contained the expected mixture of ␣-ENaC subunit forms (Fig. 4B, lane 1), and the amount of expression was similar to that of ␣-ENaC subunit forms expressed by 1393/␣-BV alone (data not shown). Similarly, cells infected with 1393/␣(⫹) BV contained high-molecular-weight aggregates in addition to the ⬃97- and 70kDa -subunit forms ( Fig. 4C, lane 1) and their total amounts are comparable to the amounts of -subunits expressed by -BV alone (data not shown). These results suggested that pVL1393 transfer vector is capable of transferring two transcription units into the baculoviral genome. We have tested the expression of ␣ and  subunits by 1393/␣(⫺)-BV in which the PHDRN␣ was oriented in
reverse orientation with respect to the -ENaC cDNA and the results are also shown in Figs. 4B and 4C. Cells infected with 1393/␣(⫺) BV did not contain any easily detectable amounts of ␣-subunit (Fig. 4B, lane 2). Although -subunit was expressed by this baculoviral infection (Fig. 4C, lane 2), its amounts were low. These results suggest that although pVL1393 transfer vector is suitable for the transfer of two cDNA’s into baculoviral genome, these transcription units should be oriented tandemly in the transfer vector to obtain their high level expression. We speculated that when two transcription units are placed in opposite orientation, unwinding of DNA during transcription from both ends would produce positive supercoiling in the middle region, thus preventing efficient transcription (23–25). We envisioned that this negative effect of positive supercoiling could be minimized by inserting a large piece of DNA in between these two oppositely oriented transcription units. To test this possibility, we have inserted a 3.0-kb ␥-ENaC cDNA (without polyhedrin promoter) in between the  and ␣ transcription units in the pVL1393/␣(⫺) transfer vector and generated pVL1393/␥␣(⫺) transfer vector. A BV termed 1393/␥␣(⫺) BV was prepared and the expression of ␣ and  ENaC subunits in Sf9 cells infected with this BV was assessed. The results show that the levels of expression of ␣-ENaC (Fig. 4B, lane 3) and -ENaC (Fig. 4C, lane 3) subunits were nearly identical to those of the ␣ and  subunits expressed by 1393/␣(⫹)-BV. These results thus suggest that the positive supercoiling effect imposed by transcription of two oppositely placed transcription units can be minimized by sufficiently increasing the distance between them. Because a polyhedrin promoter was not attached to the inserted ␥-ENaC cDNA, this subunit expression was not detected. To express ␥ subunit along with ␣ and  subunits, the ␥-ENaC cDNA carrying the polyhedrin promoter at its 5⬘-end (PHDRN␥) was isolated from pVL1393/ ␥ENaC, and NotI restriction sites were added on both ends. The resulting PHDRN␥ was then inserted at the NotI site of pVL1393/␣(⫹) transfer vector and prepared pVL1393/␥(⫹)␣(⫹) transfer vector. In this vector, all of ENaC cDNA’s are oriented tandemly (Fig. 5A). A recombinant 1393/␣(⫹)␥(⫹)-BV prepared using this vector was used to infect Sf9 cells. Immunoblots of cell membranes indicated high level expression of ␣-ENaC (lane 3, anti-␣), -ENaC (lane 3, anti-) and ␥-ENaC (lane 3, anti-␥) subunits. The amounts of these subunits were high and comparable to that of the corresponding subunits expressed by the BVs carrying a single ENaC cDNA (lanes 2 in Fig. 5B). To demonstrate directly that each Sf9 cell infected with 1393/␣(⫹)␥(⫹)-BV coexpresses ENaC subunits, the cells infected with this BV and the expressed subunits were visualized after staining with anti-␣, anti-,
SUBUNIT COEXPRESSION OF MULTIMERIC PROTEIN IN Sf9 CELLS
211
FIG. 6. Visualization of ENaC subunits in Sf9 cells infected with 1393/␣␥(⫹) (⫹)-BV. Sf9 cells were infected with 1393/␣␥(⫹) (⫹)-BV for 3 days. The cells were fixed on glass slides and incubated with anti-␣, anti-, or anti-␥ antibody, followed by Alexa-labeled anti-rabbit secondary antibody, as described under Materials and Methods. (A) The cells as observed under bright field, and (B) cells visualized with Alexa-labeled secondary antibody by indirect immunofluorescence by using scanning microscope.
FIG. 5. Construction of pVL1393 transfer vector carrying three transcription units and coexpression of ENaC subunits. The PHDRN␣ and PHDRN␥ were inserted in pVL1393/ transfer vector tandemly as described under Results (A). (B) The immunoblotting results. Membranes obtained from MDR1-BV-infected cells were run in lanes 1. Membranes obtained from cells infected with 1393/␣ (lane 2, anti-␣) 1393/ (lane 2, anti-), and 1393/␥ (lane 2, anti-␥) were run in parallel for quantitative purposes. Membranes prepared from cells infected with 1393/␥(⫹)␣(⫹)-BV were loaded in lanes 3. The blots were developed with anti-␣, anti-, and anti-␥ as indicated.
and anti-␥ ENaC antibodies in conjunction with the Alexa-coupled anti-rabbit secondary antibody, as described under Materials and Methods and the results are shown in Fig. 6. Nearly every Sf9 cell was brightly fluorescent when incubated with anti-␣, anti-, and anti-␥ antibodies, suggesting that all of the three ENaC subunits were expressed in each cell. Taken together, these results indicate that pVL1393 transfer vector can be used to transfer multiple cDNAs into the baculoviral genome for their coexpression in Sf9 cells. Functional analysis of ENaC complex. To determine the functionality of the ENaC complex, the membrane vesicles prepared from Sf9 cells infected with 1393/␣(⫹)␥(⫹)-BV were fused with planar lipid bilayers and the single channel current recordings were made. Representative channel records at several membrane voltages and the current–voltage relationship are shown in Fig. 7. The channel displayed high open probability with bursts of opening and closing transi-
tions with a mean open and close times of 1.5 and 2 s, respectively. In the presence of 1 M amiloride in cis and trans chambers, the opening and closing transi-
FIG. 7. Single channel records of Sf9 insect cell membranes expressing ENaC complex in planar lipid bilayers. Membranes prepared from Sf9 insect cells infected with 1393/␥(⫹)␣(⫹)-BV were fused to the lipid bilayer. The cis and trans chambers contained, respectively, 160 and 60 mM sodium aspartate, in the 10 mM Hepes, 0.5 mM CaCl 2 (pH 7.4) buffer. (A) Representative channel records collected at different holding potentials are shown. The open and closed states of the channel are indicated by letters O and C, respectively. (B) The behavior of the channel in the presence of 1 M amiloride in the cis and trans chambers is shown. The current– voltage plot (C) shows the unitary conductance of the channel was ⬃6 pS. The reversal potential was ⫹20 mV, which is close to the equilibrium potential for Na ⫹ (E k, ⫹23 mV).
212
RAO, MEHDI, AND STEIMLE
tions were rapid and in the presence of 10 M amiloride, the channel activity was completely blocked. The current–voltage plot shed that the unitary channel conductance of the channel was ⬃6 pS. The channels with these properties were not observed in the control Sf9 cell membrane vesicles. Thus, the long opening and closing transitions, the conductance of ⬃6 pS and sensitivity to amiloride which are characteristics of ENaC, are due to the expression of ENaC complex in Sf9 cells. DISCUSSION
The results presented in this paper mark the end to our long search for a suitable procedure for the expression of multimeric ENaC in Sf9 cells. The value of Sf9 cell–BV expression system has been widely recognized and applied to a number of soluble and membranebound proteins (5–16). As mentioned in the introduction, the most commonly used strategy for the expression of multimeric proteins is the infection of cells with a pool of BVs carrying the cDNA’s of each of the subunits required. We thus began our attempts to express the multimeric ENaC by first preparing the ␣-, -, and ␥-BVs and then infecting the Sf9 cells with these three BVs. Membrane vesicles prepared from such infected Sf9 cells exhibited very low amiloride-sensitive 22Na uptake (data not shown). The reason for this low activity of ENaC is probably due to poor assembly of fully functional ENaC. This observation is further corroborated by the immunofluorescence study on Sf9 cells infected with a pool of BVs, which indicated that about 10% of the cells were infected with a given BV and 0.1% cells were likely infected with all of the three ENaC BVs. This was further supported by the immunoblot analyses, which indicated that cells infected with multiple recombinant BVs express very low levels of ENaC subunits. Although the reason for this inefficiency is unclear at present, it may be possible that cells infected with one recombinant BV may become “immune” to further infections. It is also likely that cells infected with multiple BVs produce recombinant proteins in very small amounts that could not be detected. Although new baculoviral transfer vectors with multiple promoters are available commercially, their use in the functional expression of biologically important multimeric proteins in Sf9 cells is not reported in the literature. Whereas these vectors are useful, one caveat regarding their use is their relatively large size. To circumvent this difficulty, we constructed pAcSG3, a relatively small baculoviral transfer vector with three promoters. Insertion of ENaC subunit cDNA’s in this vector and preparation of recombinant BVs carrying one, two, and three ENaC cDNA’s were relatively straightforward. When this vector carried a single ENaC cDNA, the corresponding subunit was expressed in high amounts. However, presence of additional
cDNA’s in this vector reduced the expression of all of the cDNA’s to undetectable levels. Although a theoretical explanation for these observations is not yet found, based on the observations made with pVL1393 transfer vector in the present study, we speculate that the promoters in pAcSG3 vector are too closely positioned, preventing efficient transcription of the inserted cDNA’s. As an alternative approach to coexpress ENaC subunits, we envisioned that it is possible to insert ENaC cDNA’s as transcription units in the multiple cloning sites of a baculoviral transfer vector which can then be transferred into the baculoviral genome during homologous recombination event that take place inside the Sf9 cell. We have chosen a single polyhedrin promotercontaining pVL1393 transfer vector for this purpose based on the fact that it is a primary choice for preparing recombinant BVs in many laboratories. The -ENaC cDNA was first subcloned under the control of in-house polyhedrin promoter of pVL1393 transfer vector. When the ␣-ENaC transcription unit (PHDRN␣) was placed at the end of the -ENaC cDNA in the opposite orientation, the expression of ␣- and -ENaC subunits was greatly reduced. This observation suggests that the reduced expression could be due to the transcription of these two cDNA’s from both ends, leading to the formation of positive supercoiling in the middle region. Such positive supercoiling of DNA was previously shown to reduce the transcription (24, 25). This indicates that the transcription units are to be present in tandem to avoid the negative effects of supercoiling of DNA. This conclusion was further supported when the PHDRN␣ and -ENaC were inserted tandemly in the pVL1393 transfer vector, the expression of ␣- and -ENaC were equivalent to the their expression driven by ␣- and -BVs. Interestingly, when a 3.0-kb DNA fragment was inserted in between two oppositely oriented genes in the pVL1393/␣(⫺) transfer vector, the expression of  and ␣ subunits was again high. This suggests that when two oppositely oriented transcription units are sufficiently separated, the effects of positive supercoiling of DNA could be eliminated. The validity of these interpretations is evident from Sf9 cells infected with the 1393/␥(⫹)␣(⫹)-BV, in which the ␣-, -, and ␥-ENaC cDNA’s are present tandemly, resulted in the high yield expression of these three ENaC subunits. In summary, the results presented in this paper demonstrate that the commonly used pVL1393 transfer vector can be used to coexpress multiple proteins by inserting the required cDNA’s fused with the strong polyhedrin promoter at their 5⬘-ends in the multiple cloning sites region of the vector. Although a large number of transcription units could theoretically be inserted in the baculoviral transfer vector, a limitation appears to be the size of the transfer vector that can be
SUBUNIT COEXPRESSION OF MULTIMERIC PROTEIN IN Sf9 CELLS
constructed using the currently available molecular biological procedures. We conclude that other small baculoviral transfer vectors may also be useful in accomplishing the goal of coexpression of multimeric proteins in Sf9 insect cells. ACKNOWLEDGMENTS This work was supported by the National Institutes of Health Grant DK 51529 and the Developmental Fund, LB595 of the State of Nebraska to U.S.R. We are deeply indebted to Dr. Robert L. Rosenberg of the University of North Carolina at Chapel Hill for introducing us to the planar lipid bilayer reconstitution procedures. We thank Drs. Richard C. Boucher and J. K Stutts of the University of North Carolina at Chapel Hill, for the cDNA’s and antibodies. We also thank Shanthy L. Nuti for helpful discussions and A. S. Prakash for technical assistance.
REFERENCES 1. Studier, F. W., Rosenberg, A. H., Dunn, J. J., and Dubendorff, J. W. (1990) Methods Enzymol. 185, 60 – 89. 2. Ghosh, S., and Lowenstein, J. M. (1996) Gene 176, 249 –255. 3. Sherman, F., Fink, G. R., and Hicks, J. B. (1986) Methods in Yeast Genetics, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 4. Parekh, R. B. (1991) Curr. Opin. Biotechnol. 2, 730 –734. 5. Moss, B. (1991) Science 252, 1662–1667. 6. Lucknow, V. A., and Summers, M. D. (1988) Biotechnology 6, 47–55. 7. Volkman, L. E. (1995) Science 269, 1834. 8. Sun, T., Naini, A. A., and Miller, C. (1994) Biochemistry 33, 9992–9999. 9. Birnir, B., Tierney, M. L., Howitt, S. M., Cox, G. B., and Gage, P. W. (1992) Proc. R. Soc. London 250, 307–312. 10. Blanco, G., Xie, Z. J., and Mercer, R. W. (1993) Proc. Natl. Acad. Sci. USA 90, 1824 –1828.
213
11. DeTomaso, A. W., Xie, Z. J., Liu, G., and Mercer, R. W. (1993) J. Biol. Chem. 268, 1470 –1478. 12. Roy, P., Mikhailov, M., and Bishop, D. H. L. (1997) Gene 190, 119 –129. 13. Kost, T. A., and Condreay, J. P. (1999) Curr. Opin. Biotechnol. 10, 428 – 433. 14. Hasnain, S. E., Jain, A., Habib, S., Ghosh, S., Chatterji, U., Ramachandran, A., Das, P., Venkaiah, B., Pandey, S., Liang, B., Ranjan, A., Natarajan, K., and Azim, C. A. (1997) Gene 190, 113–118. 15. Ailor, E., and Betenbaugh, M. J. (1999) Curr. Opin. Biotechnol. 10, 142–145. 16. Chatterji, U., Ahmad, R., Venkaiah, B., and Hasnain, S. E. (1996) Gene 171, 209 –213. 17. An, K., Gillock, E. T., Sweat, J. A., Reeves, W. M., and Consigli, R. A. (1999) J. Gen. Physiol. 80, 1009 –1016. 18. Canessa, C. M., Schild, L., Buell, G., Thorens, B., Gautschi, I., Horisberger, J. D., and Rossier, B. C. (1994) Nature 367, 463– 467. 19. Rossier, B. C., Canessa, C. M., Schild, L., and Horisberger, J. D. (1994) Curr. Opin. Nephrol. Hypertens. 3, 487– 496. 20. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 21. Rao, U. S. (1995) J. Biol. Chem. 270, 6686 – 6690. 22. Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 193, 265–275. 23. Herbert, A., and Rich, A. (1996) J. Biol. Chem. 271, 11595– 11598. 24. Mizutani, M., Ura, K., and Hirose, S. (1991) Nucleic Acids Res. 19, 2907–291. 25. Liu, L. F., and Wang, J. C. (1987) Proc. Natl. Acad. Sci. USA 84, 7024 –7027. 26. Bensadoun, A., and Weinstein, D. (1976) Anal. Biochem. 70, 241–250. 27. Rosenberg, R. L., and Chen, X.-h. (1991) J. Gen. Physiol. 97, 1207–1225.
Expression of Amiloride-Sensitive Sodium Channel: A Strategy for the Coexpression of Multimeric Membrane Protein in Sf9 Insect Cells U. Subrahmanyeswara Rao, 1 A. Mehdi, and R. E. Steimle Department of Biochemistry and Molecular Biology, University of Nebraska Medical Center, Omaha, Nebraska 68198
Received February 29, 2000
The amiloride-sensitive epithelial sodium channel (ENaC) mediates Na ⴙ reabsorption in many epithelial tissues including the distal nephron, colon, lung, and secretory glands and plays an important role in pathophysiology of hypertension and cystic fibrosis. The ENaC is a multimeric integral membrane protein formed by the association of highly homologous,␣-, -, and ␥-ENaC subunits. Here we explored the Sf9 insect cell– baculovirus expression system as a source to obtain high yields of recombinant ENaC for functional and structural studies. Although this expression system is widely used, coexpression of ENaC subunits could not be accomplished by the conventional procedures. We thus developed a protocol in which the ␣and ␥-ENaC cDNA’s were first fused individually with polyhedrin promoters at their 5ⴕ-ends and then inserted in the multiple cloning sites of pVL1393 transfer vector carrying the -ENaC cDNA. Utilizing this transfer vector, a recombinant baculovirus carrying all of the three ENaC cDNA’s was prepared. Infection of Sf9 insect cells with this recombinant baculovirus resulted in the expression all of the three ENaC subunits in high yield. Planar lipid bilayer reconstitution procedure revealed the presence of ⬃6 pS sodium channels that are amiloride-sensitive. The results presented point out certain underlying rules for the expression of multiple genes in Sf9 cells, which may be useful in the expression other multimeric proteins and in the studies of protein–protein interactions as well. © 2000 Academic Press
1 To whom correspondence should be addressed at Department of Biochemistry and Molecular Biology, Campus 98-4525, University of Nebraska Medical Center, Omaha, NE 68198-4525. Fax: 402-5596650. E-mail: [email protected].
206
Development of heterologous protein expression systems such as Escherichia coli, yeast, mammalian, and Sf9 cells has revolutionized our understanding of detailed structure and function of numerous proteins (1– 7). In many instances, the expressed recombinant proteins are monomeric and function independently without the association with other proteins (8, 9). However, the scenario is different for multimeric proteins, which require coexpression of all the required subunits to elicit their functional characteristics (10, 11). Similarly, although certain proteins may independently be expressed, the biological effects of the expressed protein can only be ascertained by their interactions with other proteins. Production of biologically active multimeric recombinant proteins has been difficult and often hampered by the lack of simple procedures for simultaneous expression of all the required subunits in each of the expressing host cell. While expression of such multimeric proteins in yeast can be accomplished easily due to the availability of multiple auxotrophic yeast strains and vectors (3), the formidable and inefficient yeast cell-wall breakage procedures greatly reduce the recovery of the expressed proteins to unacceptable levels for further structural or functional analyses. Recently, the Sf9 cell– baculovirus (BV) 2 expression system has become an important tool for the largescale production of foreign proteins (12–16). Despite the fact that functional multimeric proteins were also expressed in Sf9 cells, expression of subunits of these protein complexes were interestingly accomplished by coinfection with multiple BVs each harboring the required subunit cDNA. The issues, for example, whether exposure with multiple recombinant BVs results in the infection of each Sf9 cell with all of the viruses used, and whether the expressed subunits to 2 Abbreviations used: BV, baculovirus; ENaC, epithelial sodium channel; PBS, phosphate-buffered saline.
0003-2697/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
SUBUNIT COEXPRESSION OF MULTIMERIC PROTEIN IN Sf9 CELLS
what extent are in the multimeric state, are however unclear from the literature survey. It is thus desirable to prepare a single recombinant BV harboring all of the required cDNA’s for the production of fully functional multimeric proteins in high yields. Although strategies for the transfer of multiple cDNA’s into the baculoviral genome have been formulated (12–17), such procedures for the production of biologically important functional multimeric proteins in Sf9 cells were not reported in the literature. The long-term goal of our laboratory is to understand the amiloride-sensitive epithelial sodium channel (ENaC)-mediated sodium transport. It has been established that fully functional ENaC is an integral membrane protein composed of three highly homologous subunits, termed ␣-, -, and ␥-ENaC (18, 19). To develop the Sf9 insect cell–BV expression system as a source of recombinant ENaC, we have examined several strategies to coexpress all of the three ENaC subunits in Sf9 cells. The results reported in this paper point out that coexpression of ENaC subunits in Sf9 cells is not a trivial matter, and the conventional strategy of infection of cells with multiple recombinant BVs was inefficient. Here we report a procedure that utilizes the most commonly used pVL1393 transfer vector to coexpress all of the three cDNA’s of ENaC in Sf9 cells in high amounts. The results of this study reveal that when two transcription units in the transfer vector are oppositely oriented, the expression of corresponding proteins was very low, probably due to positive supercoiling of DNA due to transcription from both ends. However, the expression is maximum when these transcription units are present tandemly in the baculoviral genome. These observations may be important in the considerations of coexpression of other multimeric proteins and in the studies of protein-protein interactions as well. MATERIALS AND METHODS
cDNA’s. The rat colon ENaC ␣-, -, and ␥-subunit cDNA’s (18, 19) were first subcloned under the control of polyhedrin promoter in the baculoviral transfer vector, pVL1393, and generated three transfer vectors, pVL1393/␣, pVL1393/ and pVL1393/␥, respectively. The ENaC cDNA’s from these vectors were used for subsequent subcloning into other baculoviral transfer vectors. Construction of baculoviral transfer vectors carrying ENaC cDNA’s and preparation of recombinant BVs. Standard molecular biological procedures of subcloning were used (20). Details of the construction of a variety of baculoviral transfer vectors carrying ENaC subunit cDNA’s are described under Results. The Sf9 cells were cotransfected with baculoviral transfer vectors and the BaculoGold DNA by the calcium phos-
207
phate coprecipitation method according to the protocol provided by the manufacturer (PharMingen, San Diego, CA) and prepared recombinant BVs. Baculoviral infections of Sf9 cells and preparation of membrane fraction. Sf9 cells were grown and infected with the recombinant BVs at a ratio of 1:5, and the total membrane fraction was prepared as described (21). SDS–PAGE and immunoblotting. Proteins in the membrane fractions were first separated on 7.5% acrylamide gels containing 0.1% sodium dodecyl sulfate and transferred onto the poly(vinylidene fluoride) membranes as described previously (21). Immunodetection of ENaC subunits was carried out by using ENaC subunit-specific polyclonal antibodies raised in rabbits in conjunction with a commercially available horseradish peroxidase-linked anti-rabbit secondary antibody (Amersham, IL). The peroxidase-labeled blots were developed by the enhanced chemiluminescence method, using the Amersham ECL Plus kit (21). Immunofluorescence microscopy. Recombinant BVinfected Sf9 cells were placed on microscope slides and fixed with neat glutaraldehyde followed by immersion in cold acetone and methanol. The slides were treated with BSA buffer (1% (w/v) bovine serum albumin dissolved in phosphate-buffered saline (PBS)) for 1 h at room temperature. The slides were then incubated with the ENaC-specific antibody diluted in PBS for 3 h, washed three times in BSA buffer, and then incubated with Alexa-coupled anti-rabbit antibody (Molecular Probes, Eugene, OR) prepared in PBS (1000:1 dilution) for 1 h at 4°C in dark. The slides were washed with PBS and fixed with antifade solution, and the Alexalabeled cells were detected under Zeiss LSM 410 confocal laser scanning microscope and the images were digitally recorded on a computer. Planar lipid bilayer reconstitution. Single channel analysis of the ENaC complex expressed by the 1393/ ␣(⫹)␥(⫹)-BV in Sf9 cell membranes was performed in planar lipid bilayers by the general procedures of Rosenberg and Chen (27) with modifications. Planar lipid bilayers were formed by painting 40 mg/ml solution of phospholipid (phosphatidylethanolamine:phosphatidylserine at a ratio of 3:1) in n-decane over a 200-m-diameter aperture separating cis and trans chambers. The cis compartment of the bilayer chamber contained 160 mM sodium aspartate, whereas, the trans chamber contained 60 mM sodium aspartate, in 10 mM Hepes, 0.5 mM CaCl 2, pH 7.4, buffer. Channel activity was monitored after the addition of Sf9 cell membranes to the cis chamber. Single channel currents collected at different voltage potentials were filtered at 100 Hz and analyzed by using pCLAMP 7.0 software (Axon Instruments).
208
RAO, MEHDI, AND STEIMLE
FIG. 1. Expression of ENaC subunits by multiple baculoviral infection. Sf9 cells were infected with a pool of ␣-,-, and ␥-BVs, and the membranes prepared were run on SDS–PAGE followed by electroblotting onto PVDF membranes (lanes 2). Membranes obtained from Sf9 cells infected with a single BV were run in parallel (lanes 1). Membranes from cells infected with BV carrying MDR1 cDNA were used as control (lanes 3). The immunoblots were developed with anti-␣, anti-, and anti-␥ antibodies. Arrows indicate the expected glycosylated and nonglycosylated subunits and their sizes.
Protein estimation. Protein in the membrane fractions was determined by the modified Lowry method, using bovine serum albumin as standard (22, 26). Materials. BaculoGold DNA and transfer vectors, pVL1393, pAcSG2, and pAcAB3 were obtained from PharMingen. The Sf9 culture media were obtained from GibcoBRL. RESULTS
Coexpression of ENaC subunits by infections with multiple recombinant BVs. To accomplish the coexpression of ␣-, -, and ␥-ENaC subunits in each Sf9 cell, the conventional procedure of infecting cells with multiple recombinant BVs was tested and the results are shown in Fig. 1. The Sf9 cells were infected with a pool of ␣-, -, and ␥-BVs and the amount of expression of each ENaC subunit in the membranes was compared with the corresponding subunit expressed in cells infected with a single BV. At least eight proteins with molecular masses in the range of ⬃97– 60 kDa were reactive with the anti-␣-ENaC antibody in membranes obtained from cells infected with ␣-BV (lane 1, anti-␣). The 97- and 80-kDa proteins are, respectively, the fully glycosylated and nonglycosylated forms of ␣-ENaC subunit, and the remaining protein bands represent proteolytic products (data not shown). Infection of cells with the pool of ␣-, -, and ␥-BVs also resulted in the expression of ⬃97- and 80-kDa ␣-ENaC subunits (lane 2, anti-␣), but the staining intensities were at least 20-fold lower than the amounts of ␣-ENaC expressed by the ␣- BV infection. Figure 1 (anti-) shows that cells infected with the -BV expressed ⬃97-kDa (glycosylated) and 70-kDa (nonglycosylated) -ENaC subunit forms (lane 1), the amounts of which were greatly reduced in cells infected with the pool of ␣-, -, and ␥-BVs (lane 2). Similarly, infection of Sf9 cells with a single ␥-BV resulted in the expression of ⬃80-kDa (glycosylated) and 75-kDa (nonglycosylated) ␥-ENaC subunits (Fig. 1, lane 1, anti-␥). However the staining intensities of these subunit forms were again greatly
reduced by coinfection of cells with the pool of ␣-, -, and ␥-BVs (Fig. 1, lane 2). We have varied the multiplicity of infection from 1:1 to 1:25 with essentially the same results (data not shown). These results suggested that coinfection of Sf9 cells with multiple BVs results in the reduced expression of each of the three ENaC subunits. To determine the reason for the reduced expression of ENaC subunits in cells infected with multiple BVs, Sf9 cells were infected with the pool of ␣-, -, and ␥-BVs and the ENaC subunits in these cells were visualized by immunofluorescence microscopy using antibodies specific to the three ENaC subunits. The results show (Fig. 2) that about 10% of the cells were fluorescent with each of the three ENaC subunit-specific antibodies, further corroborating the above immunoblotting analyses, in which the expression of these subunits was greatly reduced (Fig. 1). In contrast, nearly 100% of the cells were fluorescent when infected with any single BV (data not shown). These data suggest that cells once infected with one recombinant BV may not be further infected with any other BV. Alternatively, all of the cells were in fact infected with three ENaC BVs, but unable to produce detectable amounts of recombinant proteins. Coexpression of ENaC subunits via infection with a BV carrying three ENaC cDNA’s. To express the three ENaC subunits in each infected Sf9 cell, a BV
FIG. 2. Visualization of ENaC subunits in Sf9 cells infected with a pool of 3 ENaC BVs. Sf9 cells were infected with a pool of ␣-, -, and ␥-BVs for 3 days. The cells were fixed on glass slides and incubated with anti-␣, anti-, or anti-␥ antibody, followed by Alexa-labeled anti-rabbit secondary antibody, as described under Materials and Methods. (A) The cells as observed under bright field. The ENaC antibodies were visualized with Alexa-labeled secondary antibody by indirect immunofluorescence as observed under scanning microscope (B).
SUBUNIT COEXPRESSION OF MULTIMERIC PROTEIN IN Sf9 CELLS
FIG. 3. Construction of pAcSG3 vector as a vehicle for the expression of multiple ENaC subunits. (A–C) The construction of pAcSG3 transfer vector carrying three promoters and insertion of the three ENaC cDNA’s downstream to these promoters (see Results for details). (D) Membranes prepared from Sf9 cells infected with SG3/␣-, SG3/-, SG3/␥-, SG3/␣-, and SG3/␣␥-BVs were analyzed for the expression of ␣-, -, and ␥-ENaC subunits by immunoblotting using anti-␣, anti-, and anti-␥ ENaC antibodies. Membranes obtained from SG3/␣-BV-infected (anti-␣), SG3/-BV-infected (anti-), and SG3/␥-BV-infected (anti-␥) cells were loaded in lanes 1. Lanes 2 and 3 were loaded with membranes prepared from Sf9 cells infected with SG3/␣- and SG3/␣␥-BVs, respectively. Lanes 4 were loaded with control Sf9 cell membranes.
carrying all of the three ENaC subunit cDNA’s is necessary. Toward this end, a small baculoviral transfer vector, pAcSG3 vector containing three strong baculoviral promoters was prepared (Fig. 3) in the following way. Step 1: The commercially available ⬃10 kbp pAcAB3 baculoviral transfer vector containing three baculoviral promoters (one polyhedrin and two p10 promoters) was digested with NaeI and SnaBI and the released 1.26 kb DNA fragment that contained all of the three promoters was isolated (Fig. 3B). Step 2: A single polyhedrin promoter containing 5.5 kb baculoviral transfer vector, pAcSG2 (Fig. 3A), was digested with NaeI and SnaBI and the released 572-bp polyhedrin promoter- and the multiple cloning sites-containing fragment was discarded. Step 3: The 1.26 kb NaeI–SnaBI fragment isolated from Step 1 was ligated with the ⬃5-kb fragment isolated in Step 2 and generated a small 6.2-kb vector termed, pAcSG3, which contained one polyhedrin and two p10 promoters (Fig. 3C).
209
The ␣-ENaC cDNA was first subcloned at StuI site under the control of polyhedrin promoter, and prepared a baculoviral transfer vector, pAcSG3/␣. The -ENaC cDNA was then subcloned under the control of a p10 promoter at the SmaI site of the pAcSG3/␣ vector and generated a baculoviral vector, pAcSG3/␣. In a final step, the ␥-ENaC cDNA was subcloned under the control of the second p10 promoter at the CelII site and prepared a baculoviral vector, pAcSG3/␣␥. Using pAcSG3/␣, pAcSG3/␣, and pAcSG3/␣␥ vectors, three recombinant BVs termed SG3/␣-, SG3/␣-, and SG3/ ␣␥-BVs, which harbored, respectively, one, two, and three ENaC cDNA’s, were prepared. Total membrane fraction prepared from Sf9 cells infected with SG3/␣-, SG3/␣-, and SG3/␣␥-BVs were analyzed for the presence of ENaC subunits by immunoblotting procedure and the results are shown in Fig. 3D. Infection with SG3/␣-BV results in the expression of ␣ENaC subunit (Fig. 3, lane 1, anti-␣). Similarly, BVs prepared from the pAcSG3 vector carrying single -ENaC or ␥-ENaC cDNA, termed SG3/ (Fig. 3, lane 1, anti-) and SG3/␥ (Fig. 3, lane 1, anti-␥) allowed the expression of these subunits. However, cells infected with SG3/␣-BV (Fig. 3, lane 2) or SG3/␣␥-BV (lane 3) did not contain any detectable ␣-ENaC. There was no detectable -ENaC subunit expression in cells infected with SG3/␣- or SG3/␣␥-BVs (Fig. 3, lanes 2 and 3, anti-). Similarly, SG3/␣␥-BV infection did not result in any detectable ␥-ENaC subunit expression (Fig. 3, lane 2, anti-␥). These data suggest that although pAcSG3 vector allows high-yield expression of a single cDNA, it is not suitable for the expression of multiple cDNA’s in the Sf9 cells. High-yield expression of multiple proteins using single pVL1393 transfer vector. We have devised a new strategy to accomplish the goal of coexpression of ENaC subunits in Sf9 cells. In this procedure, the ENaC subunit cDNA’s were first ligated with polyhedrin promoters at their 5⬘-ends and generated individual transcription units which were then inserted in the multiple cloning sites of pVL1393 transfer vector. The resulting transfer vector thus contained multiple cDNA’s whose expressions were driven by polyhedrin promoters placed at their 5⬘-ends. Figure 4A shows schematically a pVL1393 transfer vector with two ENaC cDNA’s. As a first step, we have inserted the -ENaC cDNA in pVL1393 transfer vector under the control of polyhedrin promoter and prepared pVL1393/ -ENaC transfer vector. An ⬃3-kb DNA fragment comprising the entire polyhedrin promoter and the ␣-ENaC cDNA (PHDRN␣) was isolated from the pAcSG3/␣ after digestion with BamHI, and inserted at the BglII site of pVL1393/ENaC. Because BglII and BamHI cleaved ends are compatible, insertion of PHDRN␣ into pVL1393/-ENaC generated two trans-
210
RAO, MEHDI, AND STEIMLE
FIG. 4. Construction of pVL1393 transfer vector carrying two transcription units and protein expression. (A) The ␣-ENaC cDNA with polyhedrin promoter at the 5⬘-end (PHDRN␣) was prepared and then inserted downstream to the -ENaC cDNA in the pVL1393/ transfer vector. The two possible orientations of inserted PHDRN␣ are indicated. Membranes prepared from cells infected with 1393/ ␣(⫹)-BV (lanes 1), 1393/␣(⫺)-BV (lanes 2), and 1393/␥␣(⫺)-BV (lane 3) were run on SDS–PAGE followed by immunoblotting. The blots were developed with anti-␣ (B) and anti- C).
fer vectors, pVL1393//PHDRN␣(⫹) and pVL1393// PHDRN␣(⫺) in which the PHDRN␣ was inserted, respectively, in tandem and in reverse orientations, with respect to the orientation of -ENaC cDNA. The recombinant BVs prepared from these vectors are termed, respectively, 1393/␣(⫹)- and 1393/␣(⫺)-BVs. Expression of ␣- and -ENaC subunits in Sf9 cells infected with the 1393/␣(⫹)-BV was assessed by immunoblotting procedure and the results are shown in Figs. 4B and 4C. Cells infected with 1393/␣(⫹) BV contained the expected mixture of ␣-ENaC subunit forms (Fig. 4B, lane 1), and the amount of expression was similar to that of ␣-ENaC subunit forms expressed by 1393/␣-BV alone (data not shown). Similarly, cells infected with 1393/␣(⫹) BV contained high-molecular-weight aggregates in addition to the ⬃97- and 70kDa -subunit forms ( Fig. 4C, lane 1) and their total amounts are comparable to the amounts of -subunits expressed by -BV alone (data not shown). These results suggested that pVL1393 transfer vector is capable of transferring two transcription units into the baculoviral genome. We have tested the expression of ␣ and  subunits by 1393/␣(⫺)-BV in which the PHDRN␣ was oriented in
reverse orientation with respect to the -ENaC cDNA and the results are also shown in Figs. 4B and 4C. Cells infected with 1393/␣(⫺) BV did not contain any easily detectable amounts of ␣-subunit (Fig. 4B, lane 2). Although -subunit was expressed by this baculoviral infection (Fig. 4C, lane 2), its amounts were low. These results suggest that although pVL1393 transfer vector is suitable for the transfer of two cDNA’s into baculoviral genome, these transcription units should be oriented tandemly in the transfer vector to obtain their high level expression. We speculated that when two transcription units are placed in opposite orientation, unwinding of DNA during transcription from both ends would produce positive supercoiling in the middle region, thus preventing efficient transcription (23–25). We envisioned that this negative effect of positive supercoiling could be minimized by inserting a large piece of DNA in between these two oppositely oriented transcription units. To test this possibility, we have inserted a 3.0-kb ␥-ENaC cDNA (without polyhedrin promoter) in between the  and ␣ transcription units in the pVL1393/␣(⫺) transfer vector and generated pVL1393/␥␣(⫺) transfer vector. A BV termed 1393/␥␣(⫺) BV was prepared and the expression of ␣ and  ENaC subunits in Sf9 cells infected with this BV was assessed. The results show that the levels of expression of ␣-ENaC (Fig. 4B, lane 3) and -ENaC (Fig. 4C, lane 3) subunits were nearly identical to those of the ␣ and  subunits expressed by 1393/␣(⫹)-BV. These results thus suggest that the positive supercoiling effect imposed by transcription of two oppositely placed transcription units can be minimized by sufficiently increasing the distance between them. Because a polyhedrin promoter was not attached to the inserted ␥-ENaC cDNA, this subunit expression was not detected. To express ␥ subunit along with ␣ and  subunits, the ␥-ENaC cDNA carrying the polyhedrin promoter at its 5⬘-end (PHDRN␥) was isolated from pVL1393/ ␥ENaC, and NotI restriction sites were added on both ends. The resulting PHDRN␥ was then inserted at the NotI site of pVL1393/␣(⫹) transfer vector and prepared pVL1393/␥(⫹)␣(⫹) transfer vector. In this vector, all of ENaC cDNA’s are oriented tandemly (Fig. 5A). A recombinant 1393/␣(⫹)␥(⫹)-BV prepared using this vector was used to infect Sf9 cells. Immunoblots of cell membranes indicated high level expression of ␣-ENaC (lane 3, anti-␣), -ENaC (lane 3, anti-) and ␥-ENaC (lane 3, anti-␥) subunits. The amounts of these subunits were high and comparable to that of the corresponding subunits expressed by the BVs carrying a single ENaC cDNA (lanes 2 in Fig. 5B). To demonstrate directly that each Sf9 cell infected with 1393/␣(⫹)␥(⫹)-BV coexpresses ENaC subunits, the cells infected with this BV and the expressed subunits were visualized after staining with anti-␣, anti-,
SUBUNIT COEXPRESSION OF MULTIMERIC PROTEIN IN Sf9 CELLS
211
FIG. 6. Visualization of ENaC subunits in Sf9 cells infected with 1393/␣␥(⫹) (⫹)-BV. Sf9 cells were infected with 1393/␣␥(⫹) (⫹)-BV for 3 days. The cells were fixed on glass slides and incubated with anti-␣, anti-, or anti-␥ antibody, followed by Alexa-labeled anti-rabbit secondary antibody, as described under Materials and Methods. (A) The cells as observed under bright field, and (B) cells visualized with Alexa-labeled secondary antibody by indirect immunofluorescence by using scanning microscope.
FIG. 5. Construction of pVL1393 transfer vector carrying three transcription units and coexpression of ENaC subunits. The PHDRN␣ and PHDRN␥ were inserted in pVL1393/ transfer vector tandemly as described under Results (A). (B) The immunoblotting results. Membranes obtained from MDR1-BV-infected cells were run in lanes 1. Membranes obtained from cells infected with 1393/␣ (lane 2, anti-␣) 1393/ (lane 2, anti-), and 1393/␥ (lane 2, anti-␥) were run in parallel for quantitative purposes. Membranes prepared from cells infected with 1393/␥(⫹)␣(⫹)-BV were loaded in lanes 3. The blots were developed with anti-␣, anti-, and anti-␥ as indicated.
and anti-␥ ENaC antibodies in conjunction with the Alexa-coupled anti-rabbit secondary antibody, as described under Materials and Methods and the results are shown in Fig. 6. Nearly every Sf9 cell was brightly fluorescent when incubated with anti-␣, anti-, and anti-␥ antibodies, suggesting that all of the three ENaC subunits were expressed in each cell. Taken together, these results indicate that pVL1393 transfer vector can be used to transfer multiple cDNAs into the baculoviral genome for their coexpression in Sf9 cells. Functional analysis of ENaC complex. To determine the functionality of the ENaC complex, the membrane vesicles prepared from Sf9 cells infected with 1393/␣(⫹)␥(⫹)-BV were fused with planar lipid bilayers and the single channel current recordings were made. Representative channel records at several membrane voltages and the current–voltage relationship are shown in Fig. 7. The channel displayed high open probability with bursts of opening and closing transi-
tions with a mean open and close times of 1.5 and 2 s, respectively. In the presence of 1 M amiloride in cis and trans chambers, the opening and closing transi-
FIG. 7. Single channel records of Sf9 insect cell membranes expressing ENaC complex in planar lipid bilayers. Membranes prepared from Sf9 insect cells infected with 1393/␥(⫹)␣(⫹)-BV were fused to the lipid bilayer. The cis and trans chambers contained, respectively, 160 and 60 mM sodium aspartate, in the 10 mM Hepes, 0.5 mM CaCl 2 (pH 7.4) buffer. (A) Representative channel records collected at different holding potentials are shown. The open and closed states of the channel are indicated by letters O and C, respectively. (B) The behavior of the channel in the presence of 1 M amiloride in the cis and trans chambers is shown. The current– voltage plot (C) shows the unitary conductance of the channel was ⬃6 pS. The reversal potential was ⫹20 mV, which is close to the equilibrium potential for Na ⫹ (E k, ⫹23 mV).
212
RAO, MEHDI, AND STEIMLE
tions were rapid and in the presence of 10 M amiloride, the channel activity was completely blocked. The current–voltage plot shed that the unitary channel conductance of the channel was ⬃6 pS. The channels with these properties were not observed in the control Sf9 cell membrane vesicles. Thus, the long opening and closing transitions, the conductance of ⬃6 pS and sensitivity to amiloride which are characteristics of ENaC, are due to the expression of ENaC complex in Sf9 cells. DISCUSSION
The results presented in this paper mark the end to our long search for a suitable procedure for the expression of multimeric ENaC in Sf9 cells. The value of Sf9 cell–BV expression system has been widely recognized and applied to a number of soluble and membranebound proteins (5–16). As mentioned in the introduction, the most commonly used strategy for the expression of multimeric proteins is the infection of cells with a pool of BVs carrying the cDNA’s of each of the subunits required. We thus began our attempts to express the multimeric ENaC by first preparing the ␣-, -, and ␥-BVs and then infecting the Sf9 cells with these three BVs. Membrane vesicles prepared from such infected Sf9 cells exhibited very low amiloride-sensitive 22Na uptake (data not shown). The reason for this low activity of ENaC is probably due to poor assembly of fully functional ENaC. This observation is further corroborated by the immunofluorescence study on Sf9 cells infected with a pool of BVs, which indicated that about 10% of the cells were infected with a given BV and 0.1% cells were likely infected with all of the three ENaC BVs. This was further supported by the immunoblot analyses, which indicated that cells infected with multiple recombinant BVs express very low levels of ENaC subunits. Although the reason for this inefficiency is unclear at present, it may be possible that cells infected with one recombinant BV may become “immune” to further infections. It is also likely that cells infected with multiple BVs produce recombinant proteins in very small amounts that could not be detected. Although new baculoviral transfer vectors with multiple promoters are available commercially, their use in the functional expression of biologically important multimeric proteins in Sf9 cells is not reported in the literature. Whereas these vectors are useful, one caveat regarding their use is their relatively large size. To circumvent this difficulty, we constructed pAcSG3, a relatively small baculoviral transfer vector with three promoters. Insertion of ENaC subunit cDNA’s in this vector and preparation of recombinant BVs carrying one, two, and three ENaC cDNA’s were relatively straightforward. When this vector carried a single ENaC cDNA, the corresponding subunit was expressed in high amounts. However, presence of additional
cDNA’s in this vector reduced the expression of all of the cDNA’s to undetectable levels. Although a theoretical explanation for these observations is not yet found, based on the observations made with pVL1393 transfer vector in the present study, we speculate that the promoters in pAcSG3 vector are too closely positioned, preventing efficient transcription of the inserted cDNA’s. As an alternative approach to coexpress ENaC subunits, we envisioned that it is possible to insert ENaC cDNA’s as transcription units in the multiple cloning sites of a baculoviral transfer vector which can then be transferred into the baculoviral genome during homologous recombination event that take place inside the Sf9 cell. We have chosen a single polyhedrin promotercontaining pVL1393 transfer vector for this purpose based on the fact that it is a primary choice for preparing recombinant BVs in many laboratories. The -ENaC cDNA was first subcloned under the control of in-house polyhedrin promoter of pVL1393 transfer vector. When the ␣-ENaC transcription unit (PHDRN␣) was placed at the end of the -ENaC cDNA in the opposite orientation, the expression of ␣- and -ENaC subunits was greatly reduced. This observation suggests that the reduced expression could be due to the transcription of these two cDNA’s from both ends, leading to the formation of positive supercoiling in the middle region. Such positive supercoiling of DNA was previously shown to reduce the transcription (24, 25). This indicates that the transcription units are to be present in tandem to avoid the negative effects of supercoiling of DNA. This conclusion was further supported when the PHDRN␣ and -ENaC were inserted tandemly in the pVL1393 transfer vector, the expression of ␣- and -ENaC were equivalent to the their expression driven by ␣- and -BVs. Interestingly, when a 3.0-kb DNA fragment was inserted in between two oppositely oriented genes in the pVL1393/␣(⫺) transfer vector, the expression of  and ␣ subunits was again high. This suggests that when two oppositely oriented transcription units are sufficiently separated, the effects of positive supercoiling of DNA could be eliminated. The validity of these interpretations is evident from Sf9 cells infected with the 1393/␥(⫹)␣(⫹)-BV, in which the ␣-, -, and ␥-ENaC cDNA’s are present tandemly, resulted in the high yield expression of these three ENaC subunits. In summary, the results presented in this paper demonstrate that the commonly used pVL1393 transfer vector can be used to coexpress multiple proteins by inserting the required cDNA’s fused with the strong polyhedrin promoter at their 5⬘-ends in the multiple cloning sites region of the vector. Although a large number of transcription units could theoretically be inserted in the baculoviral transfer vector, a limitation appears to be the size of the transfer vector that can be
SUBUNIT COEXPRESSION OF MULTIMERIC PROTEIN IN Sf9 CELLS
constructed using the currently available molecular biological procedures. We conclude that other small baculoviral transfer vectors may also be useful in accomplishing the goal of coexpression of multimeric proteins in Sf9 insect cells. ACKNOWLEDGMENTS This work was supported by the National Institutes of Health Grant DK 51529 and the Developmental Fund, LB595 of the State of Nebraska to U.S.R. We are deeply indebted to Dr. Robert L. Rosenberg of the University of North Carolina at Chapel Hill for introducing us to the planar lipid bilayer reconstitution procedures. We thank Drs. Richard C. Boucher and J. K Stutts of the University of North Carolina at Chapel Hill, for the cDNA’s and antibodies. We also thank Shanthy L. Nuti for helpful discussions and A. S. Prakash for technical assistance.
REFERENCES 1. Studier, F. W., Rosenberg, A. H., Dunn, J. J., and Dubendorff, J. W. (1990) Methods Enzymol. 185, 60 – 89. 2. Ghosh, S., and Lowenstein, J. M. (1996) Gene 176, 249 –255. 3. Sherman, F., Fink, G. R., and Hicks, J. B. (1986) Methods in Yeast Genetics, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 4. Parekh, R. B. (1991) Curr. Opin. Biotechnol. 2, 730 –734. 5. Moss, B. (1991) Science 252, 1662–1667. 6. Lucknow, V. A., and Summers, M. D. (1988) Biotechnology 6, 47–55. 7. Volkman, L. E. (1995) Science 269, 1834. 8. Sun, T., Naini, A. A., and Miller, C. (1994) Biochemistry 33, 9992–9999. 9. Birnir, B., Tierney, M. L., Howitt, S. M., Cox, G. B., and Gage, P. W. (1992) Proc. R. Soc. London 250, 307–312. 10. Blanco, G., Xie, Z. J., and Mercer, R. W. (1993) Proc. Natl. Acad. Sci. USA 90, 1824 –1828.
213
11. DeTomaso, A. W., Xie, Z. J., Liu, G., and Mercer, R. W. (1993) J. Biol. Chem. 268, 1470 –1478. 12. Roy, P., Mikhailov, M., and Bishop, D. H. L. (1997) Gene 190, 119 –129. 13. Kost, T. A., and Condreay, J. P. (1999) Curr. Opin. Biotechnol. 10, 428 – 433. 14. Hasnain, S. E., Jain, A., Habib, S., Ghosh, S., Chatterji, U., Ramachandran, A., Das, P., Venkaiah, B., Pandey, S., Liang, B., Ranjan, A., Natarajan, K., and Azim, C. A. (1997) Gene 190, 113–118. 15. Ailor, E., and Betenbaugh, M. J. (1999) Curr. Opin. Biotechnol. 10, 142–145. 16. Chatterji, U., Ahmad, R., Venkaiah, B., and Hasnain, S. E. (1996) Gene 171, 209 –213. 17. An, K., Gillock, E. T., Sweat, J. A., Reeves, W. M., and Consigli, R. A. (1999) J. Gen. Physiol. 80, 1009 –1016. 18. Canessa, C. M., Schild, L., Buell, G., Thorens, B., Gautschi, I., Horisberger, J. D., and Rossier, B. C. (1994) Nature 367, 463– 467. 19. Rossier, B. C., Canessa, C. M., Schild, L., and Horisberger, J. D. (1994) Curr. Opin. Nephrol. Hypertens. 3, 487– 496. 20. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 21. Rao, U. S. (1995) J. Biol. Chem. 270, 6686 – 6690. 22. Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 193, 265–275. 23. Herbert, A., and Rich, A. (1996) J. Biol. Chem. 271, 11595– 11598. 24. Mizutani, M., Ura, K., and Hirose, S. (1991) Nucleic Acids Res. 19, 2907–291. 25. Liu, L. F., and Wang, J. C. (1987) Proc. Natl. Acad. Sci. USA 84, 7024 –7027. 26. Bensadoun, A., and Weinstein, D. (1976) Anal. Biochem. 70, 241–250. 27. Rosenberg, R. L., and Chen, X.-h. (1991) J. Gen. Physiol. 97, 1207–1225.